基于改进型果蝇算法的Richards模型在地铁工程沉降预测中的应用

Application of Richards Model based on Improved Drosophila Optimization Algorithm in Subsidence Prediction of Subway Engineering

针对地铁沉降具有非线性的特点以及Richards模型参数估计困难的问题,该文建立了改进型果蝇算法优化的Richards预测模型。引入变步长,并将标准粒子群算法和变步长果蝇优化算法相结合,通过粒子种群和果蝇种群对最优位置信息的相互共享来更快地找到Richards模型参数的最优值。为验证该预测模型的准确性和有效性,结合工程实例,借助C++编程与遗传算法、标准粒子群算法、果蝇算法等优化的Richards模型进行比较。程序运行结果表明,改进型果蝇算法优化的Richards模型预测精度更高,收敛速度更快,适用于地铁沉降形变预测。

In view of the nonlinear characteristics of subway subsidence and the difficult in parameters estimation of Richards model, this paper establishes a Richards prediction model optimized by improved drosophila algorithm. The optimal value of Richards model parameters can be found faster by sharing the optimal position information between particle population and drosophila population by introducing variable step size and combining the standard particle swarm algorithm and variable step drosophila optimization algorithm. In order to verify the accuracy and effectiveness of the model, combined with the engineering, the optimized Richards models by genetic algorithm, standard particle swarm algorithm and drosophila algorithm are compared by means of C++ programming. The results of the program show that the Richards model optimized by the improved drosophila algorithm has higher accuracy and faster convergence speed, which is suitable for the prediction of subway settlement deformation.